Age compensation in biometric systems using time-interval, gender and age

Abstract

Systems and methods may generate, by a computer, a voice model for an enrollee based upon a set of one or more features extracted from a first audio sample received at a first time; receive at a second time a second audio sample associated with a caller; generate a likelihood score for the second audio sample by applying the voice model associated with the enrollee on the set of features extracted from the second audio sample associated with the caller, the likelihood score indicating a likelihood that the caller is the enrollee; calibrate the likelihood score based upon a time interval from the first time to the second time and at least one of: an enrollee age at the first time and an enrollee gender; and authenticate the caller as the enrollee upon the computer determining that the likelihood score satisfies a predetermined threshold score.

What is claimed is: 
     
       1. A method of validating telecommunications comprising:
 generating, by a computer, a voice model for an enrollee based upon a set of one or more features extracted from a first audio sample received at a first time; 
 extracting, by the computer, the set of features from a second audio sample associated with a caller received at a second time; 
 generating, by the computer, a first likelihood score the caller is the enrollee by applying the voice model associated with the enrollee on the set of features extracted from the second audio sample; 
 calibrating, by the computer, the first likelihood score based upon a time interval from the first time to the second time and at least one of an enrollee age at the first time or an enrollee gender, thereby generating a second likelihood score; and 
 denying, by the computer, authentication of the caller as the enrollee in response to determining that a difference between the second likelihood score and the first likelihood score fails a predetermined threshold difference. 
 
     
     
       2. The method according to  claim 1 , wherein the predetermined threshold difference indicates a predicted change to an enrollee voice of the enrollee between the first time and the second time. 
     
     
       3. The method according to  claim 1 , further comprising authenticating, by the computer, the caller as the enrollee in response to the computer determining that the second likelihood score satisfies a predetermined threshold score. 
     
     
       4. The method according to  claim 3 , further comprising updating, by the computer, the voice model of the enrollee upon authenticating the caller as the enrollee, the voice model updated according to the set of features extracted from the second audio sample. 
     
     
       5. The method according to  claim 1 , further comprising determining, by the computer, the enrollee age at the first time based upon at least one of an age-indicator characteristic of the first audio sample and an age-indicator data point received via one or more networks from a third-party database. 
     
     
       6. The method according to  claim 1 , further comprising determining, by the computer, an approximate age of the enrollee at the second time based upon the time interval and the enrollee age at the first time,
 wherein the computer calibrates the first likelihood score to generate the second likelihood using the approximate age of the enrollee at the second time. 
 
     
     
       7. The method according to  claim 1 , further comprising:
 determining, by the computer, the enrollee gender based upon at least one of: a gender-indicator characteristic identified in the first audio sample, and a data point from a third-party database; and 
 determining, by the computer, a caller gender of the caller based upon the gender-indicator characteristic identified in the second audio sample. 
 
     
     
       8. The method according to  claim 1 , further comprising:
 generating, by a computer, an image model for the enrollee based upon a set of one or more image features extracted from a first image sample received at the first time; 
 receiving, by the computer, at the second time a second image sample associated with the caller; 
 generating, by the computer, a first image likelihood score for the second image sample by applying the image model associated with the enrollee on the set of image features extracted from the second image sample associated with the caller, the first image likelihood score indicating the likelihood that the caller is the enrollee; and 
 calibrating, by the computer, the first image likelihood score based upon the time interval from the first time to the second time and at least one of: the enrollee age at the first time and the enrollee gender, thereby generating a second image likelihood score. 
 
     
     
       9. The method according to  claim 8 , further comprising denying, by the computer, authentication of the caller as the enrollee in response to determining that an image difference between the second image likelihood score and the first image likelihood score fails an image predetermined threshold difference,
 wherein the image predetermined threshold difference indicates a second predicted change to an enrollee face of the enrollee between the first time and the second time. 
 
     
     
       10. The method according to  claim 8 , further comprising authenticating, by the computer, the caller as the enrollee upon the computer determining that the second image likelihood score satisfies a second predetermined threshold score. 
     
     
       11. A system comprising:
 a database configured to store one or more audio samples associated with one or more enrollees, the one or more audio samples including a first audio sample; and 
 a computer comprising a processor configured to:
 generate a voice model for an enrollee based upon a set of one or more features extracted from a first audio sample received at a first time; 
 extract the set of features from a second audio sample associated with a caller received at a second time; 
 generate a first likelihood score the caller is the enrollee by applying the voice model associated with the enrollee on the set of features extracted from the second audio sample; 
 calibrate the first likelihood score based upon a time interval from the first time to the second time and at least one of an enrollee age at the first time or an enrollee gender, thereby generating a second likelihood score; and 
 deny authentication of the caller as the enrollee in response to determining that a difference between the second likelihood score and the first likelihood score fails a predetermined threshold difference. 
 
 
     
     
       12. The system according to  claim 11 , wherein the predetermined threshold difference indicates a predicted change to an enrollee voice of the enrollee between the first time and the second time. 
     
     
       13. The system according to  claim 11 , wherein the computer is further configured to authenticating, by the computer, the caller as the enrollee in response to the computer determining that the second likelihood score satisfies a predetermined threshold score. 
     
     
       14. The system according to  claim 11 , wherein the computer is further configured to updating, by the computer, the voice model of the enrollee upon authenticating the caller as the enrollee, the voice model updated according to the set of features extracted from the second audio sample. 
     
     
       15. The system according to  claim 11 , wherein the computer is further configured to determine the enrollee age at the first time based upon at least one of an age-indicator characteristic of the first audio sample and an age-indicator data point received via one or more networks from a third-party database. 
     
     
       16. The system according to  claim 11 , wherein the computer is further configured to determine an approximate age of the enrollee at the second time based upon the time interval and the enrollee age at the first time, and
 wherein the computer calibrates the first likelihood score to generate the second likelihood using the approximate age of the enrollee at the second time. 
 
     
     
       17. The system according to  claim 11 , wherein the computer is further configured to:
 determine the enrollee gender based upon at least one of: a gender-indicator characteristic identified in the first audio sample, and a data point from a third-party database; and 
 determine a caller gender of the caller based upon the gender-indicator characteristic identified in the second audio sample. 
 
     
     
       18. The system according to  claim 11 , wherein the computer is further configured to:
 generate an image model for the enrollee based upon a set of one or more image features extracted from a first image sample received at the first time; 
 receive at the second time a second image sample associated with the caller; 
 generate a first image likelihood score for the second image sample by applying the image model associated with the enrollee on the set of image features extracted from the second image sample associated with the caller, the first image likelihood score indicating the likelihood that the caller is the enrollee; and 
 calibrate the first image likelihood score based upon the time interval from the first time to the second time and at least one of: the enrollee age at the first time and the enrollee gender, thereby generating a second image likelihood score. 
 
     
     
       19. The system according to  claim 18 , wherein the computer is further configured to deny authentication of the caller as the enrollee in response to determining that an image difference between the second image likelihood score and the first image likelihood score fails an image predetermined threshold difference, and
 wherein the image predetermined threshold difference indicates a second predicted change to an enrollee face of the enrollee between the first time and the second time. 
 
     
     
       20. The system according to  claim 18 , wherein the computer is further configured to authenticate the caller as the enrollee upon the computer determining that the second image likelihood score satisfies a second predetermined threshold score.